GEOGRAPHIC VARIATION IN CACTUS WREN SONGS JONATHAN L. ATWOOD and SUSANNAH B. LERMAN, Conservation Biology Program, Department of Environmental Studies, Antioch University, New England, 40 Avon Street, Keene, New Hampshire 03431-3516 (current address of Lerman: Graduate Program in Organismic & Evolutionary Biology, University of Massachu- setts Amherst, 319 Morrill Science Center South, 611 N. Pleasant Street, Amherst, Massachusetts 01003) ABSTRACT: We compared Cactus Wren (Campylorhynchus brunneicapillus) songs recorded in three regions of coastal southern California (Ventura and Los Angeles counties, Orange County, and San Diego County), Baja California, and the Sonoran and Chihuahuan deserts. On the basis of four measures of the fine structure of individual notes, songs of wrens from Baja California south of 31° N latitude were most distinct; songs of birds from the three regions of coastal southern California were similar and most like songs given by birds in Baja California. Cactus Wrens in coastal southern California are geographically isolated, morphologically different, and differ in song behavior from those in Baja California. Compared with Sonoran and Chihuahuan desert populations, Cactus Wrens in coastal southern California are geographically isolated, differ in song behavior, and occur in a unique and unusual ecological setting. These characteristics suggest that the U.S. Fish and Wildlife Service should reconsider its 1994 decision denying coastal Cactus Wrens protection under the Endangered Species Act because the population was deemed to not meet the definition of a “distinct population segment.” The Cactus Wren (Campylorhynchus brunneicapillus) is one of the most familiar and conspicuous members of the avifauna of North America’s arid regions. Distributed commonly and widely throughout the lowlands of Baja California and the Mojave, Sonoran, and Chihuahuan deserts, the species also occurs in coastal southern California and extreme northwestern Baja California (American Ornithologists’ Union 1998). Rea and Weaver (1990) reviewed the nomenclatural history and distribu- tional limits of various subspecies of the Cactus Wren in detail; these authors observed that the peninsular forms (C. b. bryanti and affinis) are so distinct morphologically from the continental populations (C. b. brunneicapillus, anthonyi, couesi, seri, and guttatus) that the two groups “look like different species.” More recently, on the basis of differences in mitochondrial DNA (mtDNA), Zink et al. (2001) similarly concluded that the peninsular and continental forms represent “evolutionarily significant unit[s] (Moritz 1994), if not phylogenetic species (Cracraft 1989).” Particular attention has recently focused on the taxonomy of Cactus Wren populations found in coastal southern California, from Ventura County south through San Diego County and into extreme northwestern Baja California. These populations are geographically separated from both the peninsular birds of Baja California and from populations in the Sonoran and Mojave deserts of California. Rea (1986) suggested that the coastal population of Cactus Wrens inhabiting southern California and extreme northwestern Baja California be recognized as a newly described subspecies, C. b. sandiegensis. Later, Rea and Weaver (1990) revised the proposed distribution (and spelling) of sandiegensis to extend from Valle de las Palmas (32° 24' N latitude) in Western Birds 38:29–46, 2007 29 GEOGRAPHIC VARIATION IN CACTUS WREN SONGS extreme northwestern Baja California only as far north as southern Orange County (San Juan Creek), California; they assigned wrens on the coastal slope of Ventura, Los Angeles, Riverside, San Bernardino, and northern Orange counties to the wide-ranging Sonoran Desert subspecies anthonyi (= couesi, sensu American Ornithologists’ Union 1957). However, Rea and Weaver (1990) commented that “some individuals [from the Los Angeles area] show some genetic influence of C. b. sandiegensis.” Although Rea and Weaver’s (1990) evaluation of sandiegensis was based on plumage characters, they also stated that “we can recognize a distinct song dialect in southern coastal wrens [sandiegensis]. The vocalizations have a slightly lower frequency and lower pitch than more northern [coastal anthonyi] and eastern [desert anthonyi] birds, and have a raspy quality not heard in adjacent populations . central and northern Orange County wrens sing the Los Angeles Basin dialect.” Sibley (2000) confused the issue by ascribing Rea and Weaver’s (1990) description of the plumage and vo- cal characteristics of sandiegensis to all Cactus Wrens in coastal southern California, including coastal populations of anthonyi. Proudfoot et al. (2000) placed sandiegensis (and, inexplicably, seri) within the peninsular group, presumably on the basis of distribution. The distribution and nomenclature of Cactus Wren subspecies is espe- cially important with regard to the coastal birds of southern California and extreme northwestern Baja California. These wrens are entirely restricted to the coastal sage scrub plant community (Rea and Weaver 1990), widely considered to be one of the most threatened habitat types in the United States as a result of intense pressure for urban development (Hanes 1976, Kirkpatrick and Hutchinson 1977, Mooney 1977, Westman 1987, O’Leary 1990). Because of these threats, in 1990 the U. S. Fish and Wildlife Service (USFWS) was petitioned by the San Diego Biodiversity Project and Palomar Audubon Society to list the San Diego Cactus Wren (C. b. sandiegensis) as an endangered species under the U.S. Endangered Species Act (ESA). The ESA uses the term “species” to include “any distinct population segment of any species of vertebrate.” That is, from a legal standpoint, the word “spe- cies” may or may not have any taxonomic connotation. The USFWS rarely uses population-level listings; when they are applied, such rulings require proof not only that a population is imperiled but also that it is both discrete and significant (U.S. Department of the Interior and U.S. Department of Commerce 1996). In response to several unpublished letters disputing the subspecific valid- ity of sandiegensis, the USWFS suggested that “Cactus Wrens residing in coastal southern California may . constitute a distinct vertebrate popula- tion segment that could qualify for listing under the [ESA]” (USFWS 1994) and expanded the subject of the original petition from sandiegensis to include all Cactus Wrens found in the coastal sage scrub of southern Cali- fornia and extreme northwestern Baja California. Thus, circumventing the question of subspecies limits, the USFWS opted instead to assess whether sandiegensis and coastal populations of anthonyi, combined, should be afforded protection under the ESA. The USFWS found that “all of the published literature on the status of coastal sage scrub vegetation types in California supports the conclusion 30 GEOGRAPHIC VARIATION IN CACTUS WREN SONGS that this plant community is one of the most depleted habitat types in the United States,” and that “numbers of Cactus Wrens residing in coastal sage scrub of coastal southern California have declined as a result . of the same factors that have . impacted” the California Gnatcatcher (Polioptila c. californica), formally designated as threatened (USFWS 1994: 45660). Nonetheless, the USFWS’s final conclusion was that coastal Cactus Wrens “do not constitute a distinct vertebrate population segment” and therefore were not “a listable entity” (USFWS 1994: 45660). The rationale supporting this decision hinged on three points (USFWS 1994: 45660). First, by broadening the petition to include all of California’s coastal Cactus Wrens, including sandiegensis and coastal populations of anthonyi, the USFWS not surprisingly concluded that “no apparent morpho- logical or other morphometric differences . distinguish coastal birds from other Cactus Wrens;” Rea and Weaver (1990:99–100) had earlier observed that “northern coastal Cactus Wrens [anthonyi, coastal populations] . are not taxonomically separable from adjacent desert anthonyi.” Second, the USFWS stated that coastal Cactus Wrens lack distinctive behavioral char- acteristics distinguishing them from desert populations. Third, the USFWS claimed that the coastal sage scrub inhabited by Cactus Wrens in coastal southern California and extreme northwestern Baja California transitions gradually into plant communities of the Baja California peninsula that are also occupied by the species, thereby proving that the wrens of coastal southern California are “not isolated in habitat.” This paper examines the conclusion that “coastal Cactus Wrens” (sensu USFWS 1994) do not represent a “listable entity” under the ESA. In par- ticular, we discuss the USFWS’s contention that coastal populations do not differ behaviorally from desert birds by comparing songs recorded throughout much of the species’ range. We also consider the USFWS’s statement that coastal Cactus Wrens are not ecologically isolated from conspecifics in the Baja California peninsula, especially in light of vocal, morphologic, and genetic differences between continental and peninsular populations. METHODS We recorded songs of 157 Cactus Wrens from 1982 to 1992 in the coastal sage scrub of southern California (Ventura and Los Angeles counties, n = 24 individuals; southern Orange County, n = 17; San Diego County, n = 17), Baja California south of 31° N latitude (n = 19), the Sonoran desert of southeastern California, southern Arizona, and northern Sonora,